The invention of photography vastly reduced the demand for painters, first by allowing a photographer to produce vast numbers of portraits quickly and cheaply, then by allowing consumers to make their own photographs. People still make paintings for money, but there is a power-law distribution of the rewards - the vast majority are amateurs or barely scrape a living, while a tiny handful of artists sell paintings for tens of millions of dollars.
There is a growing economic consensus that automation will lead to substantial polarization of the labour market. The jobs most threatened by automation are the skilled but routine jobs that characterise most of the middle class. Some of the displaced factory workers, truck drivers and clerical workers might find a good career as an artisanal baker or a professional YouTuber; I suspect that many more will end up falling into poorly paid and precarious work at the bottom of the labour market, or dropping out of the labour market entirely.
Ultimately, we're talking about people's lives and livelihoods. It's simply not good enough to assume that everything will be OK, to hand-wave away serious doubts being raised by sensible and informed people. There is a strong possibility that the wave of populism spreading across the western world is being stoked by increasing economic insecurity resulting from automation. We on HN are the people responsible for automation; it is incumbent on us to seriously examine the risks posed by it. If we are not prepared for these risks, the results could be catastrophic.
However what you haven't addressed is the impact of reducing the human labor cost/skill to producing portraits. Historically painting portraits were prohibitively expensive, likely most people never had a painting/portrait of themselves or their family. The introduction of photography enabled everyone to have photos of themselves, their families, their children growing up, their wedding. In terms of labour, we started doing things with photos that only the richest people could have afforded to do with paint. Just think of how many mall photography shops there used to be in the early 90s. In the long term we wound up with a higher quality of life by making something that used to be expensive accessible, while still maintaining a large number of jobs. Just not if you're a painter.
tl;dr, automation is bad for the individual, but good for society.
We don't see it because we're immersed in it. To that guy, everything we do is gold-plated and incredibly decadent. We are far, far beyond creating and exchanging things simply for survival.
The richest people don't even create or exchange things of any substance. A guy plays basketball, makes ten million a year. Somebody writes a best-selling book, never has to work again for life. These things would have made no sense to the overwhelming majority of people just 150 years ago.
Hell, in France of the 1950s, the majority of homes outside Paris didn't have indoor plumbing.
(Consider, as a corrolary, that a major cause of building failure is moisture-induced rot: if you want to risk-mitigate your construction and reduce a major source of failure, you'll consolidate plumbing either outside the structure, or to a subsection of it which is particularly immune to rot.)
This assumes that:
(1) jobs making "things that can't be created by automation" would be enough to replace jobs replaced by automation.
(2)people whose jobs have been automated en masse will still have money to pay for that marginally useful scarce products (or consumption by the rich/upper classes will be enough to employ them).
None of those two are safe bets. There's nothing magical about jobs, and no physical or other law that guarantees that humanity will forever, whatever the technology available, still have jobs for everybody.
It's very much possible to have a world where only 5-10% of the population is required to perform the most manual/difficult to AI jobs and all the others are automated away.
We basically have 2 historical instances where this didn't happen, and we extrapolate that this is some guarantee. But it's far from guaranteed.
The industrial revolution turned most manual and farm jobs into factory jobs because early factories weren't automated enough by the technology of the time -- not because it was "inevitable".
The service factor jobs that replaced most factory jobs when factories got more automated happened because we had the increase factory production, but no way to automate those (from accounting to burger flipping). If the majority of those get automated away from AI and software, there might be new jobs, but there's no guarantee for that.
There might be some way (through a "New Deal" style effort to create various marginally useful jobs for humans), or a huge reduction in working hours, or there might be UBI, or there will be discontent and chaos and people who are more numerous than jobs.
I hope there's enough Lamborghinis to go around then...
> The Murcielago will run you about $2000 for an oil change, $4000 for plugs and the best part $12000 for an E-Gear Clutch. The Gallardo on the other hand is more docile and similar to the Ferrari 360 Modena at just $400 for an oil change, $2000 for plugs and finally about $3500 for an E-Gear clutch. The Lamborghini maintenance ranges heavily based on the line you select and should not be taken lightly as the cost of new transmission could cost you $50,000 if a replacement is needed and lets not even talk about the engine failing.
If a new transmission costs $50,000 in parts alone. How much can you charge to put it in?
And let's not forget that polishing these cars can cost thousands of dollars per shine.
Relevant link:
http://www.secretentourage.com/lifestyle/autos/cost-of-owner...
http://www.esotericdetail.com/detailing-services/ferrari-lam...
Imagine the labor savings!
A cleaning lady is more likely to be a job than a radiologist for longer time as cleaning is actually using almost the entire human spectrum of abilities while radiologist primarily utilizes two.
That doesn't make cleaning more scarce or rare and more expensive it just means there is going to be more people who fight to get the job.
There's nothing about automation that says that thing that can't be automated become rarer than they are now.
In fact, if people whose jobs were automated turn to creating those things (and what else could they do?), then things that can't be automated will be even more plentiful (and cheaper) than they are now.
It used to be that fresh fruits were rare, sometimes even priceless. A huge number of people were required to grow, tend, harvest, transport, and sell fruit, especially in the off-months.
That's all automated now. So the price of fresh fruit is tending towards zero. All of those people are not that doing that any more. Those people are still working, just not in jobs that are as heavily automated.
There will always be something you cannot automate. Because you can't, it will be expensive. Because it is expensive, people will want it. It is not logical, but it is the way our species operates. So people do stuff, other people like it, then it's automated, then they find other stuff to do that's harder to automate.
If your thesis is that there is some sort of universal automation coming? Maybe 300-500 years out, when the Great Singularity occurs (because these two events are the same thing). But no time soon.
I will grant you that we currently have no freaking idea of what the economy in 2050 will look like. I believe this is the cool part of being mortal in our age -- things are always so goddamned different between when you're born and when you die.
It's a bit like Amdahl's Law of Parallelisation.
In some cases, that irreducible input is fairly small -- microchips and software can do a great deal (though at scale, they also represent some of the most expensive, and energy-hungry systems we've ever built).
In others ... not so much. Heating water from 0C to 100C will always take 100 calories per gram (and markedly more to convert it to steam). Casting a 22 kg iron weightlifting plate will require 22 kg of iron (and all its embedded energy). Food has its inputs, including sunlight, water, nugrients, land, topsoil, pest control, sowing, and harvesting. You might move that into a lab (I tend to doubt that), but there's still going to be some irreduceable residual.
(Actually, the story may be far worse in that we've taken a system which produces 2-8x the human labour energy inputs to one which consumes 5-10x the food energy output, largely as fossil fuels. Discounting soil and nutrient inputs.)
I strongly suspect your fruit prices are asymptotically approaching some lower bound other than zero.
I already gave you just one example of something thats not going to be more scarce, human cleaners.
How is something all humans can do, going to be cheaper even if ai never can do it?
Handmade X is expensive and rare.
People will want handmade X.
I don't have to be right all the time. I'm not saying that everything that can't be automated will automatically get more expensive. It's only necessary that some things do.
But you're worse mistaken than just in quantity: there's no mechanism to make things that can't be automated more expensive. If anything it's exactly the opposite: they'll become cheaper and more plentiful:
a) automation removes N jobs
b) people who previously did those jobs turn to whatever they can find -- and they mainly turn to do what can't be automated.
c) we now have MORE people doing what can't be automated: the people that were already doing it before automation, and the people who lost their jobs and turned to "doing things that can't be automated".
d) Abundance of workers making those things (or offering those services), drives labour costs down, fills the market with those things/services and makes things/services that can't be automated get more devalued.
I didn't ask you to be right all the time, you set that bar up yourself by claiming that things that couldn't get automated would get more rare.
I just gave you an example of something that wouldn't get more expensive even if AI couldn't do it. And now that I think of it I am struggling to find anything that will get more more scarce if AI do almost everything else. That just means more humans, more potential competition and thus less rare.
Because in a world where we've automated away most of the good jobs and the threat of no job means being destitute and on the street, your argument that house cleaners become abundant seems irrefutable.
In a world where I balance my dislike for cleaning toilets with my desire for a slightly larger share of the robot-generated bounty, it's not clear there would be an over supply of house cleaners.
This is getting rather tiresome.
Let's see. My guess is that "Things that can't be created by automation" become more scarce, hence more rare. That means their price goes up, demand increases, and more folks start making that stuff.
Why do rich people buy expensive designer cars? After all, once the model T came out, the problem of people wanting to go from one place to the other was solved.